Classification problem with imbalanced data handling methods
Budget: $30 – $250 USD
Python code of a classification problem, where 3 imbalanced datasets will be used, 4 imbalanced data handling methods (Oversampling, Undersampling, SMOTE, Proposed Criteria) , 3 classification algorithms and 5 evaluation criteria will be used. and then results show be filled out in the results sheet attached.
Evaluation criteria:
1. Matthews Correlation Coefficient
2. Recall
3. Precision
4. F-score
5. Area Under the Curve (AUC)
Classification algorithms: Naïve Bayes, Logistic Regression, Support Vector Machine
Proposed imbalanced data handling method: an algorithm that check the best mix of over, under and SMOTE with different order and difference threshold % and choses that best according to MCC measure. different thresholds: [ 0 , 15% , 30% , 50% , 65% , 80% , 90% ]
methodology attached
The 3 dataset files will be sent once project granted
Evaluation criteria:
1. Matthews Correlation Coefficient
2. Recall
3. Precision
4. F-score
5. Area Under the Curve (AUC)
Classification algorithms: Naïve Bayes, Logistic Regression, Support Vector Machine
Proposed imbalanced data handling method: an algorithm that check the best mix of over, under and SMOTE with different order and difference threshold % and choses that best according to MCC measure. different thresholds: [ 0 , 15% , 30% , 50% , 65% , 80% , 90% ]
methodology attached
The 3 dataset files will be sent once project granted